Explainable Artificial intelligence based framework for orthopedic patient classification using biomechanical features

Abstract

The benefit of using explainable AI for classifying patients based on biomechanical features is that it makes AI-based diagnoses and predictions more transparent, accountable, and clinically valuable. This results in improved patient care, improved decision-making, and a deeper comprehension of conditions. The medical field has widely embraced the application of predictive models for medical diagnosis. Various AI models are used for patient classification and illness detection in the medical diagnostic sector. This study’s main objective is to help medical practitioners anticipate osteoporosis. In our study, we have utilised machine learning algorithms to evaluate the effectiveness of each algorithm in identifying and categorizing orthopedic patients. We grouped machine learning algorithms into three cohorts, viz., solitary learning, ensemble learning, and deep learning. Each cohort is evaluated on a dataset containing 310 patients’ biomechanical features. Our experimental results indicate that ensemble AI models are superior to solitary and deep learning models. The best ensemble classifier, Extra-Tree, achieved 89% accuracy and a 96% AUC score in classifying patients suffering from orthopedics. Finally, we explained the extra-tree ensemble model predictions using the LIME and SHAP explainable artificial intelligence frameworks.

Ähnliche Arbeiten